中文摘要：

深入理解青藏高原上空大气非绝热加热三维结构,有助于揭示高原热动力效应和机械强迫效应在亚洲夏季风系统中的作用机理.然而现有的高原非绝热加热率资料存在较大不确定性.本文详细比较了NCEP和ERA40再分析资料＂残差诊断法＂计算的大气非绝热加热数据,分析两种资料所反映的高原上大气非绝热加热的时空分布特点,重点比较了二者在高原南麓的差异,并结合TRMM PR降水和潜热资料分析了差异的可能原因.研究发现两种资料之间的差异在夏季最大：ERA40在高原南麓高海拔地区所诊断的非绝热加热显著大于NCEP.ERA40大气强加热区域从高原南部山脚向北延伸、越过海拔4000位势米直至高原主体的南部;而相应NCEP大气强加热区主要位于高原南麓低海拔地区,不超过海拔4000位势米界限.上述差异不仅限于贴地层（地表感热的直接影响区域）,而在400-500hPa大气层也很显著.同时发现,ERA40所估计的夏季高原南麓降水显著大于NCEP和TRMM PR的观测,这种差异在时间、空间上都与非绝热加热的差异相吻合.这说明降水所释放的潜热是造成上述差异的主要原因.分析大气加热场和大气环流的经向垂直剖面发现,ERA40在南麓高海拔地区所诊断的大气非绝热加热可向上延伸至对流层高层-300hPa,而相应NCEP大气非绝热加热主要集中在较低大气层,相应ERA40诊断的大气垂直上升速度明显强于NCEP,200hPa的水平辐散也较强.高原南麓深对流降水及其潜热的不确定性是充分理解高原-大气相互作用的主要难点.

英文摘要：

The Tibetan Plateau（TP）, a giant elevated landscape, plays a key role in modulating the Asian summer monsoon via dynamic, thermodynamic, and mechanical lifting mechanisms. The spatial and temporal distribution of atmospheric diabatic heating（DH） is vitally important to understand the coupling of energy and moisture between the land surface and the atmosphere over TP. In this study, the residual diagnosis of DH from ERA40 and NCEP reanalysis, in combination with the precipitation and latent heating（LH） results from Tropical Rainfall Measuring Mission Precipitation Radar（TRMM PR） are investigated to reveal the uncertainties in DH over TP. Significant discrepancies of column integrated DH（CDH） between ERA40 and NCEP are found in the southern slope area of TP in summer. The CDH from ERA40 is much stronger than NCEP at high altitudes on the southern slope of TP. The area with strong ERA40 atmospheric heating extends from lower altitudes to higher altitudes exceeding 4000 m isohypse. However, in the heating diagnosis of NCEP, strong atmospheric heating is only located at altitudes lower than 4000 m isohypse. These differences are evident in DH at both the ground layer and 400–500hPa layer, revealing a significant contribution from LH because sensible heating（SH） peaks at near-surface atmosphere and LH peaks at low to middle atmosphere. As an indicator of LH, the precipitation in southern TP estimated by ERA40 is strong（10 mm/d） in the 25°–30°N area from April to October. The precipitation observations from TRMM PR and the diagnosis from NCEP diagnosis do not agree about this phenomenon. Instead, the precipitation detected by TRMM PR is mainly located in the area to the south of 25°N. The spatial pattern of overestimated precipitation by ERA40 mostly matches the DH overestimation. Based on the vertical cross-section of DH and the atmospheric circulation along 92.5°E, strong positive ERA40 DH（4 K/d） over the high altitudes of southern TP（-25°–30°N） extends from surface to h